Error in routine ktokpmq (1): k+q does not fall on k-grid

[eliephys78]

Dear H. Lee,

The 'bvec' files were not produced since the code stops before calculating them.
As I recall , the velocity matrix elements were not computed in the older version of EPW-3.0.0
As far as lambda (el-ph coupling) calculations are concerned, My calculations are running fine.

Regards

[hlee]

Dear eliephys78:

As I recall , the velocity matrix elements were not computed in the older version of EPW-3.0.0
As far as lambda (el-ph coupling) calculations are concerned, My calculations are running fine.

You are right.
vme is used in transport module and adaptive smearing (triggered by degaussw=0). It seems that both of them don't apply to your case.

Sincerely,

H. Lee

[cmzhang]

Dear all,

I met the same problem with QE-6.8. The code stops at:

Use zone-centred Wigner-Seitz cells
Number of WS vectors for electrons 601
Number of WS vectors for phonons 157
Number of WS vectors for electron-phonon 157
Maximum number of cores for efficient parallelization 2826
Results may improve by using use_ws == .TRUE.

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
Error in routine ktokpmq (1):
k+q does not fall on k-grid
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

stopping ...

After locating and modifying the subroutine ktokpmq in kfold.f90 as suggested above, I checked the values of variables "xxk", "xx" in:

Code: Select all

    !
    ! now add the phonon wavevector and check that k+q falls again on the k grid
    !
    xxk = xxk + DBLE(sign) * xxq
    !
    xx = xxk(1) * nkc1
    yy = xxk(2) * nkc2
    zz = xxk(3) * nkc3
    in_the_list = ABS(xx - NINT(xx)) <= eps5 .AND. &
                  ABS(yy - NINT(yy)) <= eps5 .AND. &
                  ABS(zz - NINT(zz)) <= eps5
    IF (.NOT. in_the_list) CALL errore('ktokpmq', 'k+q does not fall on k-grid', 1)
    !

It seems the xxk is slightly away from desirable grid point value. For example, xxk is 0.08333288940037246, which is slightly smaller than 1/12 (I used a 24x24x1 grid). When multiplied by nkc1=24, the error is magnified and exceeds eps5. By replacing eps5 by a larger cutoff eps4, the code runs well so far.

Best,
Chenmu

[hlee]

Dear Chenmu:

Thank you for the inspection and for providing the solution.

Sincerely,

H. Lee